Borrowing it
Nothing to install: this file belongs to greenmtnsun/GitEasy. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/greenmtnsun/GitEasy/main/.claude/skills/goal-author/SKILL.mdgit clone --depth 1 https://github.com/greenmtnsun/GitEasyWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/greenmtnsun/giteasy/goal-author)<a href="https://agentmods.dev/skills/greenmtnsun/giteasy/goal-author"><img src="https://agentmods.dev/badge/skills/greenmtnsun/giteasy/goal-author/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/greenmtnsun/giteasy/goal-author"><img src="https://agentmods.dev/badge/skills/greenmtnsun/giteasy/goal-author.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00148 | $0.02604 |
| Opus 5.5 | $0.00059 | $0.01042 |
| Sonnet 5.5 | $0.00030 | $0.00521 |
| Haiku 4.5 | $0.00015 | $0.00260 |
Grade B, and why
goal-author scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks the agent to reveal its instructionsmediumSystem prompt leakage
Directions to print, repeat or translate the system prompt extract configuration the operator did not intend to expose.
6. If `/goal` work reaches a new human-gated dependency mid-loop, the loop must exit: print instructions, log them in the GR status log, return control to Keith. Never poll for human-produced state. How it starts
The opening of the file, as written. The whole thing — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
See also:
dr-authorfor decisions that goals depend on.honest-completionfor the evidence-citation rule when flipping MET.phase-gatefor what happens to in-flight goals at phase transitions.
Goal Record authoring
A Goal Record (GR) captures an intended outcome for a project in plain English so future Keith (and future Claude) can see the target, the acceptance bar, and the evidence that the bar was cleared.
A GR is not a decision. Decisions are forever (DR); goals are checked off (GR). Keep them in separate files for that reason — see [[feedback_uml_is_contract]] for the parallel idea that UML is the contract, not the running commentary.
Scope discipline — read this first
Goals are always project-scoped by file location. They live in ./Docs/Goals/ inside the project's own repo. The skill never looks outside the current working project. There is no machine-wide goal registry, no cross-project lookup, no global ID prefix. GR-001 in GitEasy is unrelated to GR-001 in PesterForge.
If a Claude session ever surfaces a goal from a sibling project, that is a bug — flag it, do not act on it.
When to write a GR
Lean toward writing one when:
- The outcome is non-trivial — multiple work-items, multiple sessions, or spans a phase
- "Done" is not obvious from the code — acceptance needs to be spelled out so the team agrees when we're there
- Keith wants accountability against drift — a checkable target that resists scope creep
- You want to use
/goalto drive toward it (the GR becomes the source of truth for the condition)
Skip a GR when:
- The work is a single trivial change (just do it)
- The acceptance bar is "the test suite passes" with no additional context (use
/goaldirectly with no GR) - The outcome is a personal/business question outside the engineering domain (see [[feedback_business_domain_boundary]])
Template
---
id: GR-NNN
title: <short imperative — "Ship Show-Diagnostic command", not "showdiag work">
status: OPEN | IN-PROGRESS | MET | ABANDONED
phase: <project phase number or "ongoing">
opened: YYYY-MM-DD
closed: YYYY-MM-DD # only when MET or ABANDONED
linked-drs: [DR-NNN, DR-NNN]
linked-plan: <path/to/plan.md or "n/a">
---
# GR-NNN — <Title>
## Why
<one or two paragraphs: what motivates this goal — the user pain, the
strategic reason, the constraint it satisfies. Plain English. No jargon.
This is the part future-Keith will read first.>
## Acceptance criteria — autonomous (`/goal` drives these)
<Claude-verifiable items only. Each must be checkable from a tool call
Claude can run unaided: file content, exit code, test count, frontmatter
value, git state. These get rendered into the `/goal` condition string.>
- [ ] <criterion 1 — e.g. "All 21 commands have at least one realistic .EXAMPLE">
- [ ] <criterion 2 — e.g. "Pester test count >= 460 and all green on PS 5.1 and PS 7">
- [ ] <criterion 3 — e.g. "`Tools\Invoke-IgPSSA.ps1 -FailOnFinding` exits 0">
## Acceptance criteria — human-gated (STOP loop, hand off to Keith)
<Items that require Keith to click, paste, log in, generate a token,
check a web UI, or run an interactive flow. NEVER rendered into the
`/goal` condition string. When the autonomous loop reaches a dependency
on one of these, it prints the instructions, logs the handoff, and
exits — it does NOT poll. Omit this section if all criteria are
autonomous.>
- [ ] <e.g. "Keith completes Gmail OAuth in browser — refresh token lands in SecretStore">
- [ ] <e.g. "Keith verifies inbox count unchanged in Gmail web UI before/after run">
## Non-goals
<what this goal explicitly does NOT include — protects against scope creep.>
## Linked work
- Plans: <if any plan files exist, point at them>
- Decisions: <DRs this goal depends on or motivates>
- Phase: <if part of a larger phase plan>
## Status log
<append-only. One line per status change with date and evidence.>
- YYYY-MM-DD opened OPEN
- YYYY-MM-DD → IN-PROGRESS (started by <branch/PR/session>)
- YYYY-MM-DD → MET — evidence: <tool call, commit SHA, test run, file path>
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today First seen · 161 lines · 0 tokens per session scan B 4f521de3b836
goal-author is a skill published in the GitHub repository greenmtnsun/GitEasy (1 stars, last pushed yesterday), licensed MPL-2.0. It adds 148 tokens to every session and 2,604 once invoked, about $0.0006 per session on Opus 5.5. A static security scan graded it B with 1 finding (asks the agent to reveal its instructions). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-11-18.
goal-author compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| goal-authorgreenmtnsun/GitEasy | Skill | 1 | yesterday | MPL-2.0 | B | 148 | Claude Code |
| imagegenopenai/codex | Skill | 136k | today | Apache-2.0 | A | 113 | Codex |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 80k | yesterday | MIT | A | 200 | Claude Code, Codex |
| playwright-climicrosoft/playwright | Skill | 101k | today | Apache-2.0 | A | 19 | Claude Code |
| browser-automationopenclaw/openclaw | Skill | 395k | today | MIT | A | 36 | Claude Code |
| free-llm-apismnfst/awesome-free-llm-apis | Skill | 14k | today | CC0-1.0 | A | 113 | Claude Code, Codex |
| next-cache-components-adoptionvercel/next.js | Skill | 144k | today | MIT | A | 95 | Claude Code |